The measurement of magnetic properties of tissues in the human body is fast becoming a key element in studying disease with magnetic resonance imaging (MRI) and in molecular imaging. The magnetic properties of a subvoxel object usually cannot be determined from a given magnitude image in MRI. Although the magnetic susceptibility difference between an object and its surrounding tissue leads to a signal loss in the nearby tissue, the combination of magnitude and phase images can be used to quantify the magnetic property of any given object in MR images. A novel inverse method using this feature will be fully developed in this proposal, especially for a few voxel or subvoxel objects. This method requires the use of complex MRI data and naturally accounts for the partial volume effect, dephasing effect (i.e., signal loss), and the phase aliasing effect. Compared to other methods, this method does not require any a priori information of the object of interest. The uncertainty of the method depends on the signal-to-noise ratio and resolution of images. Three specific aims are proposed in this research. The first aim is to determine the magnetic moment of a spherical object and an infinitely long cylindrical object. The former represents a 3D problem while the latter represents a 2D problem. Fundamental electromagnetism guarantees that the magnetic moment of any small object in 3D can be well approximated by the magnetic moment of a sphere. Both simulations and phantom experiments will be conducted to validate the method and investigate the uncertainty of the method. The second aim is to apply the method on existing human and animal images. This is to demonstrate the feasibility of the method on practical images. The third aim is to resolve the magnetic susceptibility and volume of the object individually. Since it is obvious that the volume of an object much less than a voxel such as a nanoparticle in images cannot be determined, it is important to study the limitation of the proposed method. This work could have a major impact on studying aging or molecular imaging through the use of nanoparticles. Evidence indicates that 2-amyloid plaque is related to Alzheimer's disease. Animal studies have shown that amyloid deposits lead to hypointense spots at an early stage of Alzheimer's disease. Quantification of the property of each individual dark spot may become a predictor for monitoring the progression of the disease. Currently, no non-invasive tool other than number counting is available for accurately quantifying microbleeds or hypointense spots in the brain. Nanoparticles have been widely used in MRI for tagging cells in molecular imaging. A reliable in vivo method is needed to quantify the concentration of nanoparticles labeling cells that interact with diseased tissue. This method can do that with numerous identical nanoparticles. In summary, this research has the potential to contribute to the better diagnosis and function of disease in MRI. PUBLIC HEALTH RELEVANCE: The proposed method may become a predictor of Alzheimer's disease in the long run. It may become a useful tool for the design of personalized medicine. The benefit of this research to the public health is obvious.